INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

anya

-0.75

review

-0.74

Reviewer

-0.71

OVER

-0.71

levels

-0.71

adder

-0.70

 âĨĳ

-0.69

zai

-0.68

doms

-0.68

cation

-0.66

POSITIVE LOGITS

bidden

0.84

 Paradise

0.71

aughtered

0.71

llor

0.70

 Andromeda

0.67

 Trident

0.65

 Juno

0.65

romeda

0.65

auga

0.63

 Allaah

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.